AI Infrastructure

2026 年 Jul 29 日
AI engineering

PJM to cut data center power during shortages starting 2027

PJM will cut power to data centers over 50 MW during shortages starting June 2027, compensating them under a demand response program.

2026 年 Jul 29 日
AI engineering

LFM2.5-Encoders: Fast Long-Context Inference on CPU

Liquid AI released open-weight encoders that handle 8,192 tokens with CPU inference roughly 3.7× faster than ModernBERT-base at long context.

2026 年 Jul 28 日
AI engineering

Recursive Superintelligence signs $410M compute deal with AWS

Recursive Superintelligence signs a $410 million compute deal with AWS, funding self-improving AI systems with most of its budget going to compute over headcount.

2026 年 Jul 28 日
Tools

Microsoft launches cybersecurity model MAI-Cyber-1-Flash and Perception platform

Microsoft launched its first cybersecurity model MAI-Cyber-1-Flash and the Perception platform, using agentic teams to automate security workflows.

2026 年 Jul 28 日
Business

SAP BDC Connect for BigQuery Now GA 90

SAP and Google Cloud launch SAP Business Data Cloud Connect for BigQuery, offering zero-copy, bidirectional data sharing between SAP and Google Cloud's big data ecosystem.

2026 年 Jul 27 日
AI engineering

Are brain waves the next unlock for physical AI?

Encord is testing brain wave headsets to generate physical training data for robots, addressing the scarcity of real-world data for humanoid and warehouse robotics.

2026 年 Jul 26 日
AI engineering

Synthetic Data and Pre-Training at Scale: poolside’s Lessons

Poolside's synthetic data pipeline pairs templates with supplementary context, while the training team crashes on weight mismatch to catch silent failures at scale during pre-training.

2026 年 Jul 26 日
AI engineering

How one power line failure exposed a growing AI data center grid problem

A power line failure near D.C. caused 3 GW of data centers to disconnect from the PJM grid simultaneously, flickering lights and highlighting a growing grid stability problem.

2026 年 Jul 26 日
AI engineering

Co-operative time-slicing for RL post-training in llm-d

Co-operative time-slicing lets RL jobs interleave on shared GPUs, raising duty cycles from 40% to 70% without hurting convergence, by checkpointing idle phases to DRAM and swapping in active ones.